Judging by the present state-of-the-art in the field of theory and philosophy of history it seems fair to say that a return to some basic definitions from earlier iterations of longstanding disciplinary debate might be timely. This appears to be particularly important in relation to the question of what it means for events to be constructed or, indeed, fictional (or “fictioned”)-even if that part of the overall debate has continued in one form or another for well over half a century already, and especially forcefully so in its “linguistic-turn” mode. For me, a usual course of action to approach this question would be to look specifically at Hayden White and so-called narrative theory of history. The problem with going this route, however, is that the narrative theory of history approach after White remains a contested perspective on these matters even today. So, I am hoping it might be helpful instead to proceed by building on several (somewhat) more mainstream, but currently largely ignored, philosophical and theoretical, as well as history-specific, points and proposals, all to some extent converging on the idea that, notoriously, “events are identified under a description.” Ultimately this leads to largely the same conclusions as those of my usual narrative theory suspects but draws on quite distinct positions and arguments from thinkers like Elizabeth Anscombe (obviously) through R.G. Collingwood, Reinhart Koselleck, Richard Rorty, John Zammito, and Barbara Herrnstein Smith to Paul Roth.-

To begin, three key points:

1) The question of construction (and “fictioning”) differs critically from the general one concerning reference that we encounter in philosophy of language. The philosophical touchstone for debates about reference since the beginning of the twentieth century has been proper names; debates vacillate between treating names as a type of ostensive act (for example, rigid designators) or treating them as linguistically complex (for example, disguised descriptions). This has no relevant connection to the question of what it is, specifically, that historians do in their work-to what happens when we try to relate or explain some congeries of past facts as an event, for instance, because the philosophical shape of these referential issues simply fails to touch the key questions that demarcate theory and philosophy of history as a distinct pursuit. Philosophers of language ask how to assess how different linguistic elements contribute to evaluating the truth conditions of a statement; debates in theory and philosophy of history post-White turn on assessing competing narratives. Answers to the former question do not answer the latter, and querying reference is a much more fundamental question (or more precisely: problematic) than that of how we do things with language when producing (re)presentations.

Thus, for pragmatic and disciplinary reasons alone, the question of reference is not what we should dwell on when investigating history and historical events. Instead, the compromise best made in theorizing history is to accept that the “facts” with which historians work are sufficiently well established to permit talk about a real past (these facts are often termed “singular existential statements,” even if that formulation is somewhat problematic; for example: “an Israeli military unit attacked the village of Qibya in October 1953”.² Notably, doing this seems more than justified pragmatically by the extent to which historians’ professional training focuses on the “research side” of their work. Historians are experts at establishing historical facts.

2) When we try to extend this same pragmatic compromise from these singular statements to the level of “the event,” things have conventionally become more challenging, however. One (I would claim: minority) approach is to think that events exist in the world as simple objects of some kind, there to be “found” like anything else. That they have some form, substance, and perhaps even meaning of their own-something that defines and fixes them. Call this “event essentialism.” And this essentialist assumption appears to be the main premise for attacks on constructivism or “narrativism.”³ For example, many of the contributions to a recent book arguing for the “poverty of anti-realism” appear to equate the event with any and every change of state along the following lines: “Historical events are then dispositions and changes of dispositions of the constituents of reality in space and time”.⁴ The essentialists presume to know (in advance of inquiry? As a result of inquiry? One is not told) what the constituents of reality must be. Theirs is a view from nowhere.

The inherence of events seems a challenging position to defend, however, since pointing to them in the world (even phenomenologically to present-day events) is an exceedingly difficult task. For it demands, inter alia, that one knows how to carve those events independently of our discursive practices. What happens once they have been carved out in this way-constructed but also “ontologized,” as it were-is a different question, and the two should not be conflated. Arguably-and fairly evidently- events as discursive constructs can be seen to exist “out there” in the world. But, even then, attaching them to specific singular existential statements (to a defined and finite set of those) is most often unfeasible. What, for instance, are the singular statements that need to be included for an account to be one of the French Revolution?-a relatively standard example. What facts need to be included in the construction of an event for the construction to be able to sufficiently constitute and demarcate it? Or, taking another classic, related example (adapting from R.G. Collingwood referring to life in London), how would one go about giving an account of “yesterday in Helsinki”?⁵ What is a sufficient or appropriate account and what is not? How do we decide what is relevant and what irrelevant? This makes plain how different are the questions asked by the theorist of history as opposed to a philosopher of language.

The obvious answer appears to be that whatever is included or excluded is defined not by the “event,” or even the complex “object,” but by (the intentions and purpose of) the account in marking that event. The account constructs the entity but does absolutely nothing, and makes no claims, on the level of the ontic, and a general agnosticism regarding actuality is something that needs to be observed throughout what follows. The problem is thus not that the referential compromise necessarily made by choice of the historical genre suddenly disappears; it is, rather, that we are judging something very different here, and validation does not work identically with complex constructions.

3) In trying to decide the relevance of including or appealing to specific singular statements, resorting to cause and effect also falls short. Although ostensible “causality” works wonders in linguistic constructions, the value of real-world causal connections is far more limited than sometimes intuitively assumed. This is all the more so when the things to be connected are far away from each other and embedded in complex situations-with real-world phenomena clearly being situated in exceedingly more complex settings than most practical and helpful examples of causality. After all, a “causal chain” can be retrospectively plotted from (almost?) anything today all the way back to the most fundamental “events,” yet doing so will provide little or no explanatory or interpretive content. To hopefully not belabor the obvious: my sitting here typing now can be causally traced to the first use of fire by humans, to metal production, or to the invention of the automobile and consequent need for garages for IT start-ups-or to anything else one can think of since causal links point and spread forward and tracing them the other way makes little sense.

So, what would any of that causal connecting mean? In contrast to this kind of weak “explanation” through causal links, meaningful connections and events are created through stronger, formative discursive practices of meaning giving. One now-standard example of these kinds of stronger meaning-making processes can be found in Arthur Danto’s idea of “narrative sentences” such as “the Thirty Years’ War began in 1618” and the additional challenges posed by the retrospective emplotment which this very simple-seeming statement already contains. Of course, these kinds of examples also problematize the idea of singular existential statements when more closely investigated, but, when focused on issues relating to referential forms like history specifically, that problem needs to remain bracketed.

Importantly, a fuller understanding of causality also helps remind why the dynamics of denying (“vetoing”) historical interpretations is possible with factual evidence (at its most fundamental à la Reinhart Koselleck: “The sources provide control over what might not be stated”⁶), whereas “proving” them is not. The fact that the process of verification can only be similarly unidirectional is yet another matter that should by now be clear in any philosophical and theoretical debates about history (or indeed about any form of referential representation) but does not seem to be.

As noted, and while much of this may appear obvious to those working in philosophy or literary studies, for example, many current attempts at theorizing history writing are in no way au fait with these preliminary issues. Fortunately, there are also some particularly insightful discussions to lean on, Paul Roth’s The Philosophical Structure of Historical Explanation⁷ providing one of the more recent ones. Here, Roth offers (among other useful formulations) what he aptly terms “the nondetachability thesis”-the idea that an event cannot be detached from a description of it. As he writes, “events explained by histories exist qua events only as constructions of those histories. […] Not, of course, because the world it depicts does not exist [NB!]; rather, the narrative selectively orders materials. […] A narrative constructs both the explanans and the explanandum”.⁸ Or, in a more direct formulation (and one immediately also connecting to my title here): “Historical events exist only as events under a description”.⁹ (Or, a few pages later, again: “historical events ‘exist’ only as part of some narrative or other”)¹⁰

In connection with this idea of nondetachability, Roth-leaning heavily on Danto and Louis Mink-presents two other theses, the three together forming his definition of historical narrative. This three-part formulation of the “defining features of historical narratives” looks deceptively simple despite its sophistication: “the nondetachability of conclusions, the nonstandardization of events explained, and the nonaggregativity of narrative explanations”.¹¹ As I see it, these bring concretely to bear four major, much discussed theoretical and philosophical issues, central also to concrete historical practice: the inevitability of imposing meanings, the complexity of social reality, the challenges of natural language-use and the impossibility of any universal history.

Although my focus with these definitions now is slightly different to Roth’s (who wants to chart how historical narrative may count as explanation), the same idea of events always existing only “under a description” follows quite naturally also from the three basic, interrelated points raised; since, (1) if meaningful collections and congeries of facts are harder (in practice) to naturalize than individual singular statements, (2) if events do not have a concrete existence “out there” to guide our final mapping of them and the selection of relevant individual statements, and (3) if causal links are not especially helpful in overcoming these difficulties, then it already seems painfully clear that practices of linguistic construction and meaning-making are the only way to even begin to think about the challenges involved in dealing with historical events, let alone their nature.¹²

So, after these preliminary considerations, what does thinking of events as established “under a description” (as construed broadly, with respect to the various uses of that formulation)¹³ commit us to? And what kinds of caveats should be registered?

The main hurdle in many debates now seems to stem from the idea or intuition that there must still be some more concrete way for events to “exist” than this constructivist one. By suggesting that there is not, we supposedly underwrite a dangerous and extreme antirealism regarding the existence of the world in general-or at least this is what many of the defenders of these kinds of critical views claim. But that is not what any of those seriously defending linguistic construction propose.

Although Roth’s claims may on first sight look extreme, he emphatically qualifies that attention here is on how “events explained by histories exist qua events,” not, say, on how they “exist” as unconstituted collections of facts in past reality. Hence critiques starting out from a perspective according to which events are an essential part of reality obviously speak past his considerations. Importantly, and as already seen above, Roth is careful not to overstate this claim, however, explicitly explaining that this in no way implies a denial of the existence of reality. In similar manner, even White and Keith Jenkins-so often the main poster-boys for antirealism in history-are vocal in denying exaggerated readings of their positions. White, for his part, makes the same pragmatic admission that I emphasize above: “The reality of the past is a given, it is an enabling presupposition of historical enquiry” (White 2005, 148). Likewise, presenting what is, ultimately, simply an antifoundationalist position, Jenkins writes: “I take as my originary axiom the existence of matter; of materiality, of ‘actuality.’ [… But] it seems apparent that the actuality of ‘existence’ skips free of every (definitive) anthropomorphism”¹⁴

In this light, and to understand current “realist” intuitions and objections better, it may help to return to more extreme presentations of the idea that accounts construct reality. The argument for a form of antirealism most suggestive of the irrelevance of reality for historical research was, to my mind, presented early in the debates by Frank Ankersmit in Narrative Logic:

The Renaissance is nothing more and nothing less than what individual historians tell us that it is. Therefore, an account of the Renaissance is quite different from e.g. a description of a physical object: whatever we decide to write down in our historiography on the Renaissance cannot but be true of our Renaissance. We cannot misdescribe the Renaissance (because there is no such thing), while it is not hard to misdescribe chairs or automobiles. At most we can defy existing historiographical customs, but to a certain extent historians are even obliged to defy such customs.¹⁵

While this way of thinking provocatively pushed the boundaries of the antirealist argument, it also introduced some concerning suggestions and may have created needless resistance. Of course, as historians, “our” the Renaissance-if it is intended to be commensurate with the historical knowledge and historiographical discussions indicated by the name-still needs to somehow involve the culture and arts of 15th- or 16th-century Europe rather than the resurgence of historical fiction in the past 30 years, for example. Further, given the definitional and pragmatic constraints regarding what counts as history, it appears equally clear that we can quite easily “misdescribe” individual building-blocks, as it were, of this Renaissance.¹⁶ In this connection, we need to be reminded of the distinction between truth and validity. Otherwise, we are unwittingly led back to dated intuitions relying on some expectation of correspondence, like Ankersmit in this instance, even if his position at this point seems to have been supported by a tension between somewhat contradictory beliefs: although he leans strongly on correspondence with the emphasis on “misdescription” here, his chosen vocabulary of the “account” of the constructed object versus a “description” of the physical one points toward arguments for assessing validity, an issue he only mentions once in passing.

To underscore this: the challenge of complex constructions should not be seen in terms of success in evaluating them with respect to their truthfulness. Not even when we compromise and agree that their individual particulars need to be evaluable in such terms for us to have practices of referentially committed writing like history at all. The challenge lies rather with the ways in which we agree to assess their validity. Given all the debates regarding these issues, it should be evident that this cannot be done in terms of correspondence, however (re)branded or (re)described.¹⁷ Yet an ideal of correspondence appears to be the persistent intuition feeding into any insistence that the difference between the Renaissance and chairs is best to be presented in terms of their amenability to misdescription. But doing away with correspondence does not amount to denying reference, only limiting its utility for validating these constructions.

The difficulty introduced by this kind of extreme provocation is that it obscures the continuity of the most critical issues from the case of the Renaissance to that of “this chair”: pertinent is not the voluminousness and current absence of the former or the physicality and presence of the latter, nor is “referentiality” the main concern, pace both Ankersmit and Zammito in their exchange. Rather, the crucial difference is in the amount and extent of linguistic evaluation and construction that is being engaged. This is the stumbling block for more recent debates too: what needs to be recognized is the fact that even if our representational practices naturalize and ontologize a complex construction like “the Renaissance” or “the 1948 war,” for example, this does not justify thinking of it as an entity-an “event,” say, that can be pointed to directly and unproblematically. Hence, rather than being swayed only by assumptions about the determining nature of reality here, we would at the very least need arguments and evidence for any opposing view. Without it, taking the idea of “not found” to the extreme of “not existing” only serves to polarize the discussion and leads to the kinds of overstatements that reproduce age-old discussions from the history of philosophy.

So, what can we do to avoid falling into these traps? To better address the issues, the validity of an interpretation needs to be distinguished from the truthfulness of the facts employed. This should be evident from the mere possibility of “revisionist” histories alone. On this point, one can only echo Ankersmit: for judging what constitutes an event or object, indeed, how that event or object is constituted, appeals can only be made to “historiographical tradition” and historians’ professional practices and societal pressures and responsibilities, to the why of the construction. But for critics then to jump from this recognition of construction to cries of wolf regarding belief in the nonexistence of reality and rejection of factual evidence looks disingenuous.

To avoid such extremes, the difference between truth and validity needs rather to be understood as a function of the introduction of (e)valuation(s), not through reference or correspondence. To add that missing aspect to the extremes from the claim discussed above: can we misdescribe a chair as a “good” chair or a “bad” chair, a “comfortable” or an “uncomfortable” one? Is the goodness or comfort of the chair “found” or “constructed”? And does its possible constructedness somehow then suggest an inability to refer to the same thing? Or a nonexistence? Does valuation then mean that we are talking about different objects? Is my “good chair” in our shared office in the morning not the same as your “bad chair” in the afternoon? Allowing one valuation to lead us to think that we are not dealing with the same reality as that mentioned in the other seems an undesirable outcome.

As I read it, one attempt intended at a mediating position between events existing to be found and their being constructed was formulated by Ankersmit some time later,¹⁸ in the shape of a thin linguistic “skin” that brings no additional “content of the form” to a representation, to also immediately introduce the alternative understanding here in Whitean terms. But this too suggests that some interpretation (or presentation of an event qua event) could somehow be “right.” Which still begs the core antirealist or antifoundationalist question of “Right, how?” Empirically? Morally? Aesthetically? In some determinable way? Is such a thin skin indeed conceivable without a tropological function or content to it, despite its linguistic nature? For me, it remains hard to see how that could be, even only in relation to the empirical; in relation to the ethical and the aesthetic, it would further entail very specific metaphysical beliefs.¹⁹

In this light, to try to define an event by what happened rather than by the significance given to those happenings-by its political or other impact-seems unsustainable. Yet the possibility of something being somehow both constructed and “existing” appears to be the recurring hurdle in these debates despite the fact that no-one is arguing against the existence of reality. (Remember Roth’s qualification: “Not, of course, because the world it depicts does not exist…”). That is to say, no-one seriously maintains that historical descriptions or constructions of “the Renaissance,” for example, pragmatically attach to nothing that existed. Likewise, it would not be useful to claim that the terms “The 1948 War,” “The War of Independence,” “the catastrophe of 1948,” and “Al-Nakbah,” for instance, do not relate to the same reality, even if the valuations brought to bear in their formulation are incommensurable.

For an event to exist as an event only “under description” in such ways is, of course, not only a peculiarity of Roth or Mink or Danto, but rather a well-rehearsed and common theme for quite diverse thinkers, which brings attention back to the compromise about philosophy versus history and, more generally, theory versus practice-a discussion that should be more familiar from recent “continental” philosophy and pragmatism than from the stricter “analytic” tradition. Consider, for example, Sartre’s “in the world,” or Derrida’s declaration that philosophical thinking can only take us so far and societal issues need to be tackled by bracketing fundamental philosophical doubts. Or consider, indeed, any formulation of antifoundationalism and its consequences.²⁰

Naturally then, when considering representations from the point of view of valuation and validity, the question of which aspects and whose perspective, whose story, we concentrate on becomes the crucial issue; as with “the formation of the State of Israel” as opposed to “the displacement of over 750,000 Palestinians from their own homeland,” for instance,²¹ and the resulting creation of a ‘“safe sanctuary” for world Jewry’²² or “a life of misery, deprivation and desperation” for the Palestinian refugees.²³ Judgements here cannot usefully be viewed as being a matter of reference or correspondence and nor can their validity be assessed on the basis of the language used, from the passive and seemingly objective-“the Arabs went from being a majority to a minority, considerably impacting on majority-minority relations in the state to this day”²⁴ -to the openly emotive: “The Palestinians in the occupied lands, who were once the majority in their native land, became an undesired and oppressed minority and were under the mercy of military decrees and emergency laws that remained in effect until 1966”.²⁵

Instead, and more fundamentally, judgements are made on the basis of how we choose to value the related privilege or suffering: for the State of Israel, “the economic, security and social issues […] were daunting” and “the war exhausted the economy of the young country, while the threat posed by the Arab countries and the Palestinians to the security of the fledgling state did not abate”.²⁶ Meanwhile, “The Palestinian refugees in these countries lived near the border areas, hoping to return […]. They found themselves homeless, wandering aimlessly, dispersed in every direction, and without any protection”.²⁷

This is particularly so since all this valuation takes place while the “facts” are seldom in question: both sides present commensurable names, dates, places, and numbers, to mention but a few, and even on occasion speak in terms of the “deportation” and “exile” of “refugees” as well as of the “massacres” committed by one side and “border tensions” and “incidents” caused as well as “infiltration” attempts made by the other. Thus, neither of these histories can be simply “vetoed” on that basic level. And nor is there reason to contest many of their interpretations; both agree that “[t]he aim of the 1948 war was to […] take control of the land and populate it with Jews,” even if the other wants to emphasize that it was also “to empty the land of its Arab population by all possible means”.²⁸ Consequently, if we are looking for the best “fit” among these kinds of competing narratives, who else could decide that except whoever that “we” happens to be?

So, where easy intuitions may suggest some natural match between one “event” and one particular colligatory concept, such desires need to be examined. Examples of overlapping “events” (and not simply nested ones) are not hard to find-something that, again, should drive home the point of essential constructedness and availability to redescription; think of “genocide in the Americas” and “the triumph of capitalism,” for instance, or-perhaps particularly today-of “the formation of the State of Israel” and “the end of the British empire.” And, obviously, overlaps grow along with interpretive or explanatory scope. Similarly, care needs to be taken with the related intuition that there are certain singular statements that need to be included in any or every representation of a recognized event. However, the fact that all such statements and ideas need to be seen (pragmatically) to “refer” is not to say that we have a commensurate reality in which agreement may be sought. Incommensurability of perspectives is based on valuation, not “truth.”

Thus, although, practical considerations return us to the necessary compromise accepted by every referential form-in the case of history, to what the historian’s most fundamental professional commitments are built on-there are important caveats. In all these respects, representations are flawed by nature, and this is the core insight that finally needs to be brought to fully bear on thinking history too. It is the unwillingness to accept this “impaired” nature of referential representation that appears to lie at the core of too many of the debates and which continuously resuscitates misgivings about the linguistic nature of meaning-making. The key challenge seems to be with the recognition of the linguistic condition alongside a simultaneous, lingering need for definitive and exhaustive representation, or at least something “solid” to fall back on, a desire to have things both ways at once, which necessarily leads to contradiction. Yet giving up on certainties, the desire for which inspires the need for representation to do more than it can, is the core lesson of antirealism and antifoundationalism. Certainty is not needed for valuation-indeed, valuations and judgements are needed specifically in relation to what we do not know. What is more, even if some level of certainty of things in the world could be achieved-the problem that all referential discourse must bracket to exist in the first place-those things can never furnish us with meanings.